3D-OSEM versus FORE + OSEM: Optimal Reconstruction Algorithm for FDG PET with a Short Acquisition Time

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Radiochemistry

Background:

• Short acquisition times in fluorodeoxyglucose (FDG) positron emission tomography (PET) are desirable for patient comfort and throughput.
• Reconstruction algorithms significantly impact image quality and diagnostic accuracy in low-dose PET imaging.

Purpose of the Study:

• To compare the performance of Fourier rebinning (FORE) + ordered subsets expectation-maximization (OSEM) and 3D-OSEM reconstruction algorithms for FDG PET imaging with short acquisition times.
• To determine the optimal algorithm for balancing image quality and lesion detectability in accelerated FDG PET scans.

Main Methods:

• Phantom studies utilized spheres of varying sizes and radioactivity ratios with acquisition times from 15 to 180 seconds.
• Clinical evaluation involved mid-abdominal FDG PET images from 19 patients, assessed visually and quantitatively (SUVmax, CV).
• Image data were reconstructed using both FORE+OSEM and 3D-OSEM algorithms for comparative analysis.

Main Results:

• FORE+OSEM yielded higher coefficients of variation (CV), indicating poorer image quality, but maintained relatively constant maximum standardized uptake values (SUVmax).
• 3D-OSEM provided better image quality (lower CV) but significantly underestimated SUVmax, with small lesions becoming obscured at short scan times.
• Phantom and clinical results showed similar trends, with FORE+OSEM demonstrating clearer FDG uptake despite lower overall image quality.

Conclusions:

• FORE+OSEM reconstruction results in lower image quality compared to 3D-OSEM for short-duration FDG PET scans.
• Despite poorer image quality, FORE+OSEM reconstruction is more effective for detecting lesions due to clearer visualization of FDG uptake in accelerated scans.